From 8745ca5efb2aad0b6591d9b8cd48573ea49c929d Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 9 Oct 2025 10:48:09 -0400 Subject: drm/amdgpu: fix initialization of doorbell array for detect and hang Initialized doorbells should be set to invalid rather than 0 to prevent driver from over counting hung doorbells since it checks against the invalid value to begin with. Signed-off-by: Jonathan Kim Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 5bf9be073cdd..30e1fb510600 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -409,7 +409,7 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, return -EINVAL; /* Clear the doorbell array before detection */ - memset(adev->mes.hung_queue_db_array_cpu_addr, 0, + memset(adev->mes.hung_queue_db_array_cpu_addr, AMDGPU_MES_INVALID_DB_OFFSET, adev->mes.hung_queue_db_array_size * sizeof(u32)); input.queue_type = queue_type; input.detect_only = detect_only; -- cgit v1.2.3 From 0ef930e1faca6418316e5b9a3b4d1f6ae9e5b240 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 9 Oct 2025 11:28:19 -0400 Subject: drm/amdgpu: fix hung reset queue array memory allocation By design the MES will return an array result that is twice the number of hung doorbells it can report. i.e. if up k reported doorbells are supported, then the second half of the array, also of length k, holds the HQD information (type/queue/pipe) where queue 1 corresponds to index 0 and k, queue 2 corresponds to index 1 and k + 1 etc ... The driver will use the HDQ info to target queue/pipe reset for hardware scheduled user compute queues. Signed-off-by: Jonathan Kim Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 30e1fb510600..94973018f761 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -420,12 +420,17 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, dev_err(adev->dev, "failed to detect and reset\n"); } else { *hung_db_num = 0; - for (i = 0; i < adev->mes.hung_queue_db_array_size; i++) { + for (i = 0; i < adev->mes.hung_queue_hqd_info_offset; i++) { if (db_array[i] != AMDGPU_MES_INVALID_DB_OFFSET) { hung_db_array[i] = db_array[i]; *hung_db_num += 1; } } + + /* + * TODO: return HQD info for MES scheduled user compute queue reset cases + * stored in hung_db_array hqd info offset to full array size + */ } return r; -- cgit v1.2.3 From 277bb0f83e98261018ddd82b7ab8154bb9b93237 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 5 Jun 2025 10:18:37 -0400 Subject: drm/amdgpu: enable suspend/resume all for gfx 12 Suspend/resume all gangs has been available for GFX12 for a while now so enable it. Signed-off-by: Jonathan Kim Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c') diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 94973018f761..4883adcfbb4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -691,14 +691,11 @@ out: bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev) { uint32_t mes_rev = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK; - bool is_supported = false; - if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && - amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && - mes_rev >= 0x63) - is_supported = true; - - return is_supported; + return ((amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && + amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && + mes_rev >= 0x63) || + amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0)); } /* Fix me -- node_id is used to identify the correct MES instances in the future */ -- cgit v1.2.3